The only nuance here is that we use view() to reshape our tensors from the LSTM output to be the correct size for input into our fully connected layer.
確定！ 回上一頁
